MIT's Moral Machine wants you to decide who dies in a self-driving car accident | Hacker News Reader

MIT's Moral Machine wants you to decide who dies in a self-driving car accident

thenextweb.com

9 pointsaburan288 comments

I understand the discomfort that people have with the notion that an autonomous car will make such life and death decisions. However, I feel like it's obviously better for such decisions to be made ahead of time in a thoughtful, collaborative manner than by one human driver in less than a second when something goes wrong.

The truth however, is that the decision will end up being made by some opaque neural net that's been trained by the actions of other human drivers in relevant situations. It'll just "react" the same way that we do. Coding specific logic into such edge cases is probably unrealistic.

As far as tech goes, this is mostly a pointless exercise.

The car is supposed to know it's in a low-speed urban environment where there may be zebra crossings, so it accelerates accordingly and always has time to break.

If a pedestrian makes an illegal crossing, the car should try to protect its occupants first, which is the most natural response for a human driver. Not to mention autonomous cars will be more likely to turn these unavoidable accidents into avoidable ones.

One can entertain all sorts of situations, but so far I haven't found one without an answer. It always boils down to not including a "split-second morality system".

This isn't a question for morals, it's obviously a question of property damage and insurance premiums. Anyone suggesting otherwise simply misses the point of what makes our great wonderful 'modern' society actually run day to day.

Either way, the car owner is responsible, not the programmer.

How can the car owner be responsible when they have made no decision (and had no opportunity to do so).

It boils down to property rights and contracts.

The decided to buy the car, they decided to operate it, they decided where to go and when.

That's only in a libertarian dystopia with "Caveat Emptor" as its motto and no contract in place to guarantee safety as long as the car works properly.

In the real world, it depends on the situation.

In this particular situation, if the car was driving so fast it does not have time to break for, the fault most likely falls to the manufacturer for writing software that defaults to reckless driving in an urban environment, and inmediately after some of the fault might fall to the city or the inspectors that approved the manufacturer if it is concluded that the approval was given either knowing the software was like this, or failing to account for it in their testing procedure.

If there was a mapping error, it depends on the map provider's license to the car manufacturer.

All other situations (car breaking down due to lack of maintenance, breaking down unexpectedly, pedestrian making an illegal move, etc,...) are already handled by current law (obligatory "some exceptions may apply").

I agree that in our current situation voluntary interaction is considered harmful.

But even in your description, it still boils down to Property Rights and Contacts (not described in those terms however). It may be that the programmers or business owners will be forced into contracts that hold them liable. IMO, that would be unfortunate.

Reply on news.ycombinator.com